With the recent release of Linux Standard Base 2.0 by the Free Standards Group, attention has once again turned to preventing the fragmentation of Linux distributions. LSB 2.0 is intended to enable software vendors to write an application once and have it install and run identically on many Linux distributions. Customers, in turn, would be able to choose from a variety of Linux distributions without fear of incompatibility.
Probably the most important new feature in LSB 2.0 is support for C++, a key requirement for many enterprise application vendors. However, like many other aspects of LSB, this feature has caused controversy in the Linux and open-source community. The C++ compliance specifications in LSB 2.0 are based on a version of GCC 3.3, but many in the open-source community consider that version old and would prefer to see it based on the newer, more capable GCC 3.4.
Creating a base standard that commercial vendors will be happy with always means choosing older, more stable technologies over more capable but untested ones. Standards such as LSB will always be a step or two behind the cutting edge, for good reason. Open-source developers should accept that.
Another source of friction is the fact that GCC (GNU Compiler Collection) is a GNU program. Richard Stallman, founder of the GNU Project, has said that GNU maintainers cannot cooperate with LSB. This kind of infighting is common in the Linux and open-source community and, while regrettable, isnt a serious threat to the future of LSB, in our view.
LSB wont go very far, however, unless it gets the full backing of the big vendors. Most key Linux distributors and major hardware vendors such as Dell, HP, AMD and Intel have openly voiced their support for LSB.
Although core enterprise application vendors such as BEA, SAP and Oracle also support the standard, their endorsements have been muted so far. For example, very little mention of Oracle is found on the Free Standards Group Web site, and little mention of LSB is found on Oracles Web site. In fact, only one application vendor, Covalent, has formally announced support for Version 2.0.
Enterprise applications are the biggest cause by far of fragmentation in the Linux market. Getting these to run on most Linux distributions is, essentially, the reason LSB exists. If major enterprise application vendors do not show enthusiasm about LSB, questions will continue about whether it has a chance to succeed. We hope that the reticence of the large vendors is due to their taking extra care in ensuring the compatibility of their applications and not due to an underlying ambivalence.
Theres no question in our mind that LSB 2.0 represents the best chance to prevent Linux fragmentation. Customers and vendors alike have plenty at stake in the existence of a robust, widely supported standard. We urge all parties, enterprise application vendors especially, to back LSB 2.0 wholeheartedly.
Let us know what you think at [email protected].